CSV Schema Inference

Overview

You send CSV data and we read a bounded sample of its rows to describe each column: the name, the type we think it holds, whether any cell is empty, how many distinct values we saw, and a few real examples. Ask for it and we also suggest an allowed-value list for columns with few distinct values. This is a well-founded guess from the rows we read, not a guarantee about rows we have not seen, so treat the result as a starting point and check data with schema.validate_csv_schema. Accepts file input via file_ref — no need to extract file content first.

Opening an unknown export and eyeballing it does not scale, and guessing the column types is how imports break at three in the morning. You get the shape of the file in one call.

When to use it

Scope

Non-goals

• We never check data against a schema — that is schema.validate_csv_schema
• We never clean, reshape, or rewrite your CSV — that is data.csv_normalize
• We never guess the delimiter for you; tell us if it is not a comma

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

File limits

Constraints this skill applies to your input and generated files. The “Limited by” column shows the static source of each cap; none of these can be raised by upgrading your plan.

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"schema.infer_csv_schema","input":{"csv":"name,age\nAlice,30"}}'

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("schema.infer_csv_schema", {"csv":"name,age,email,joined\nAlice,30,[email protected],2024-01-15\nBob,25,,2024-03-20"});

Errors

Other limits

Limitations

• We never check data against a schema — that is schema.validate_csv_schema
• We never clean, reshape, or rewrite your CSV — that is data.csv_normalize
• We never guess the delimiter for you; tell us if it is not a comma
• Checking a CSV against rules you already have (schema.validate_csv_schema)
• Cleaning or reshaping the data (data.csv_normalize)
• Profiling JSON, XML, or any other non-tabular format

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "schema.infer_csv_schema",
  "input": {
    "csv": "name,age,email,joined\nAlice,30,[email protected],2024-01-15\nBob,25,,2024-03-20"
  }
}

Examples

{
  "csv": "name,age,email,joined\nAlice,30,[email protected],2024-01-15\nBob,25,,2024-03-20"
}

{
  "warnings": [],
  "sampleSize": 2,
  "rowCount": 2,
  "columnCount": 4,
  "columns": [
    {
      "nullCount": 0,
      "nullable": false,
      "sampleValues": [
        "Alice",
        "Bob"
      ],
      "distinctCount": 2,
      "inferredType": "string",
      "name": "name"
    },
    {
      "nullCount": 0,
      "nullable": false,
      "sampleValues": [
        "30",
        "25"
      ],
      "distinctCount": 2,
      "inferredType": "integer",
      "name": "age"
    },
    {
      "nullCount": 1,
      "nullable": true,
      "sampleValues": [
        "[email protected]"
      ],
      "distinctCount": 1,
      "inferredType": "string",
      "name": "email"
    },
    {
      "nullCount": 0,
      "nullable": false,
      "sampleValues": [
        "2024-01-15",
        "2024-03-20"
      ],
      "distinctCount": 2,
      "inferredType": "date",
      "name": "joined"
    }
  ]
}

{
  "csv": "status\nactive\ninactive\npending",
  "detectEnums": true
}

{
  "warnings": [],
  "sampleSize": 3,
  "rowCount": 3,
  "columnCount": 1,
  "columns": [
    {
      "inferredType": "string",
      "distinctCount": 3,
      "suggestedEnum": [
        "active",
        "inactive",
        "pending"
      ],
      "name": "status",
      "nullCount": 0,
      "nullable": false,
      "sampleValues": [
        "active",
        "inactive",
        "pending"
      ]
    }
  ]
}

Recommended use cases

• Find out what an unfamiliar CSV export actually contains before you process it
• Draft the column rules you will later enforce with schema.validate_csv_schema
• Document a dataset for someone who has never opened it

Avoid for

• Checking a CSV against rules you already have (schema.validate_csv_schema)
• Cleaning or reshaping the data (data.csv_normalize)
• Profiling JSON, XML, or any other non-tabular format

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags